diff mbox series

[net] net: stmmac: No need to calculate speed divider when offload is disabled

Message ID 20240614081916.764761-1-xiaolei.wang@windriver.com (mailing list archive)
State New, archived
Headers show
Series [net] net: stmmac: No need to calculate speed divider when offload is disabled | expand

Commit Message

xiaolei wang June 14, 2024, 8:19 a.m. UTC 
commit be27b8965297 ("net: stmmac: replace priv->speed with
the portTransmitRate from the tc-cbs parameters") introduced
a problem. When deleting, it prompts "Invalid portTransmitRate
0 (idleSlope - sendSlope)" and exits. Add judgment on cbs.enable.
Only when offload is enabled, speed divider needs to be calculated.

Fixes: be27b8965297 ("net: stmmac: replace priv->speed with the portTransmitRate from the tc-cbs parameters")
Signed-off-by: Xiaolei Wang <xiaolei.wang@windriver.com>
---
 .../net/ethernet/stmicro/stmmac/stmmac_tc.c   | 38 ++++++++++---------
 1 file changed, 20 insertions(+), 18 deletions(-)

Comments

Simon Horman June 15, 2024, 2:47 p.m. UTC | #1 
On Fri, Jun 14, 2024 at 04:19:16PM +0800, Xiaolei Wang wrote:
> commit be27b8965297 ("net: stmmac: replace priv->speed with
> the portTransmitRate from the tc-cbs parameters") introduced
> a problem. When deleting, it prompts "Invalid portTransmitRate
> 0 (idleSlope - sendSlope)" and exits. Add judgment on cbs.enable.
> Only when offload is enabled, speed divider needs to be calculated.
> 
> Fixes: be27b8965297 ("net: stmmac: replace priv->speed with the portTransmitRate from the tc-cbs parameters")
> Signed-off-by: Xiaolei Wang <xiaolei.wang@windriver.com>
> ---
>  .../net/ethernet/stmicro/stmmac/stmmac_tc.c   | 38 ++++++++++---------
>  1 file changed, 20 insertions(+), 18 deletions(-)
> 
> diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_tc.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_tc.c
> index 1562fbdd0a04..b0fd2d6e525e 100644
> --- a/drivers/net/ethernet/stmicro/stmmac/stmmac_tc.c
> +++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_tc.c
> @@ -358,24 +358,26 @@ static int tc_setup_cbs(struct stmmac_priv *priv,
>  
>  	port_transmit_rate_kbps = qopt->idleslope - qopt->sendslope;
>  
> -	/* Port Transmit Rate and Speed Divider */
> -	switch (div_s64(port_transmit_rate_kbps, 1000)) {
> -	case SPEED_10000:
> -	case SPEED_5000:
> -		ptr = 32;
> -		break;
> -	case SPEED_2500:
> -	case SPEED_1000:
> -		ptr = 8;
> -		break;
> -	case SPEED_100:
> -		ptr = 4;
> -		break;
> -	default:
> -		netdev_err(priv->dev,
> -			   "Invalid portTransmitRate %lld (idleSlope - sendSlope)\n",
> -			   port_transmit_rate_kbps);
> -		return -EINVAL;
> +	if (qopt->enable) {
> +		/* Port Transmit Rate and Speed Divider */
> +		switch (div_s64(port_transmit_rate_kbps, 1000)) {
> +		case SPEED_10000:
> +		case SPEED_5000:
> +			ptr = 32;
> +			break;
> +		case SPEED_2500:
> +		case SPEED_1000:
> +			ptr = 8;
> +			break;
> +		case SPEED_100:
> +			ptr = 4;
> +			break;
> +		default:
> +			netdev_err(priv->dev,
> +				   "Invalid portTransmitRate %lld (idleSlope - sendSlope)\n",
> +				   port_transmit_rate_kbps);
> +			return -EINVAL;
> +		}
>  	}
>  	mode_to_use = priv->plat->tx_queues_cfg[queue].mode_to_use;

Hi Xiaolei Wang,

The code following this function looks like this:

	if (mode_to_use == MTL_QUEUE_DCB && qopt->enable) {
		ret = stmmac_dma_qmode(priv, priv->ioaddr, queue, MTL_QUEUE_AVB);
		if (ret)
			return ret;
		priv->plat->tx_queues_cfg[queue].mode_to_use = MTL_QUEUE_AVB;
	} else if (!qopt->enable) {
		ret = stmmac_dma_qmode(priv, priv->ioaddr, queue,
				       MTL_QUEUE_DCB);
		if (ret)
			return ret;
		priv->plat->tx_queues_cfg[queue].mode_to_use = MTL_QUEUE_DCB;
	}

	/* Final adjustments for HW */
	value = div_s64(qopt->idleslope * 1024ll * ptr, port_transmit_rate_kbps);
	priv->plat->tx_queues_cfg[queue].idle_slope = value & GENMASK(31, 0);

	value = div_s64(-qopt->sendslope * 1024ll * ptr, port_transmit_rate_kbps);
	priv->plat->tx_queues_cfg[queue].send_slope = value & GENMASK(31, 0);

And the div_s64() lines above appear to use
ptr uninitialised in the !qopt->enable case.

Flagged by Smatch.
xiaolei wang June 16, 2024, 1:15 a.m. UTC | #2 
On 6/15/24 22:47, Simon Horman wrote:
> CAUTION: This email comes from a non Wind River email account!
> Do not click links or open attachments unless you recognize the sender and know the content is safe.
>
> On Fri, Jun 14, 2024 at 04:19:16PM +0800, Xiaolei Wang wrote:
>> commit be27b8965297 ("net: stmmac: replace priv->speed with
>> the portTransmitRate from the tc-cbs parameters") introduced
>> a problem. When deleting, it prompts "Invalid portTransmitRate
>> 0 (idleSlope - sendSlope)" and exits. Add judgment on cbs.enable.
>> Only when offload is enabled, speed divider needs to be calculated.
>>
>> Fixes: be27b8965297 ("net: stmmac: replace priv->speed with the portTransmitRate from the tc-cbs parameters")
>> Signed-off-by: Xiaolei Wang <xiaolei.wang@windriver.com>
>> ---
>>   .../net/ethernet/stmicro/stmmac/stmmac_tc.c   | 38 ++++++++++---------
>>   1 file changed, 20 insertions(+), 18 deletions(-)
>>
>> diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_tc.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_tc.c
>> index 1562fbdd0a04..b0fd2d6e525e 100644
>> --- a/drivers/net/ethernet/stmicro/stmmac/stmmac_tc.c
>> +++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_tc.c
>> @@ -358,24 +358,26 @@ static int tc_setup_cbs(struct stmmac_priv *priv,
>>
>>        port_transmit_rate_kbps = qopt->idleslope - qopt->sendslope;
>>
>> -     /* Port Transmit Rate and Speed Divider */
>> -     switch (div_s64(port_transmit_rate_kbps, 1000)) {
>> -     case SPEED_10000:
>> -     case SPEED_5000:
>> -             ptr = 32;
>> -             break;
>> -     case SPEED_2500:
>> -     case SPEED_1000:
>> -             ptr = 8;
>> -             break;
>> -     case SPEED_100:
>> -             ptr = 4;
>> -             break;
>> -     default:
>> -             netdev_err(priv->dev,
>> -                        "Invalid portTransmitRate %lld (idleSlope - sendSlope)\n",
>> -                        port_transmit_rate_kbps);
>> -             return -EINVAL;
>> +     if (qopt->enable) {
>> +             /* Port Transmit Rate and Speed Divider */
>> +             switch (div_s64(port_transmit_rate_kbps, 1000)) {
>> +             case SPEED_10000:
>> +             case SPEED_5000:
>> +                     ptr = 32;
>> +                     break;
>> +             case SPEED_2500:
>> +             case SPEED_1000:
>> +                     ptr = 8;
>> +                     break;
>> +             case SPEED_100:
>> +                     ptr = 4;
>> +                     break;
>> +             default:
>> +                     netdev_err(priv->dev,
>> +                                "Invalid portTransmitRate %lld (idleSlope - sendSlope)\n",
>> +                                port_transmit_rate_kbps);
>> +                     return -EINVAL;
>> +             }
>>        }
>>        mode_to_use = priv->plat->tx_queues_cfg[queue].mode_to_use;
> Hi Xiaolei Wang,
>
> The code following this function looks like this:
>
>          if (mode_to_use == MTL_QUEUE_DCB && qopt->enable) {
>                  ret = stmmac_dma_qmode(priv, priv->ioaddr, queue, MTL_QUEUE_AVB);
>                  if (ret)
>                          return ret;
>                  priv->plat->tx_queues_cfg[queue].mode_to_use = MTL_QUEUE_AVB;
>          } else if (!qopt->enable) {
>                  ret = stmmac_dma_qmode(priv, priv->ioaddr, queue,
>                                         MTL_QUEUE_DCB);
>                  if (ret)
>                          return ret;
>                  priv->plat->tx_queues_cfg[queue].mode_to_use = MTL_QUEUE_DCB;
>          }
>
>          /* Final adjustments for HW */
>          value = div_s64(qopt->idleslope * 1024ll * ptr, port_transmit_rate_kbps);
>          priv->plat->tx_queues_cfg[queue].idle_slope = value & GENMASK(31, 0);
>
>          value = div_s64(-qopt->sendslope * 1024ll * ptr, port_transmit_rate_kbps);
>          priv->plat->tx_queues_cfg[queue].send_slope = value & GENMASK(31, 0);
>
> And the div_s64() lines above appear to use
> ptr uninitialised in the !qopt->enable case.

Oh, when deleting the configuration, idleslope and sendslope are both 0, 
do you mean we also need to set ptr to 0?

thanks

xiaolei

>
> Flagged by Smatch.
>
> --
> pw-bot: changes-requested
Simon Horman June 17, 2024, 10:29 a.m. UTC | #3 
On Sun, Jun 16, 2024 at 09:15:05AM +0800, xiaolei wang wrote:
> 
> On 6/15/24 22:47, Simon Horman wrote:
> > On Fri, Jun 14, 2024 at 04:19:16PM +0800, Xiaolei Wang wrote:

...

> >          /* Final adjustments for HW */
> >          value = div_s64(qopt->idleslope * 1024ll * ptr, port_transmit_rate_kbps);
> >          priv->plat->tx_queues_cfg[queue].idle_slope = value & GENMASK(31, 0);
> > 
> >          value = div_s64(-qopt->sendslope * 1024ll * ptr, port_transmit_rate_kbps);
> >          priv->plat->tx_queues_cfg[queue].send_slope = value & GENMASK(31, 0);
> > 
> > And the div_s64() lines above appear to use
> > ptr uninitialised in the !qopt->enable case.
> 
> Oh, when deleting the configuration, idleslope and sendslope are both 0, do
> you mean we also need to set ptr to 0?

Understood, if idleslope and sendslope are 0, then ptr could be set to any
value and the result would be the same.  And, based on my limited
understanding, 0 does not seem to be a bad choice.

My point is that ptr shouldn't be uninitialised at this point.

...
xiaolei wang June 18, 2024, 2:31 a.m. UTC | #4 
On 6/17/24 18:29, Simon Horman wrote:
> CAUTION: This email comes from a non Wind River email account!
> Do not click links or open attachments unless you recognize the sender and know the content is safe.
>
> On Sun, Jun 16, 2024 at 09:15:05AM +0800, xiaolei wang wrote:
>> On 6/15/24 22:47, Simon Horman wrote:
>>> On Fri, Jun 14, 2024 at 04:19:16PM +0800, Xiaolei Wang wrote:
> ...
>
>>>           /* Final adjustments for HW */
>>>           value = div_s64(qopt->idleslope * 1024ll * ptr, port_transmit_rate_kbps);
>>>           priv->plat->tx_queues_cfg[queue].idle_slope = value & GENMASK(31, 0);
>>>
>>>           value = div_s64(-qopt->sendslope * 1024ll * ptr, port_transmit_rate_kbps);
>>>           priv->plat->tx_queues_cfg[queue].send_slope = value & GENMASK(31, 0);
>>>
>>> And the div_s64() lines above appear to use
>>> ptr uninitialised in the !qopt->enable case.
>> Oh, when deleting the configuration, idleslope and sendslope are both 0, do
>> you mean we also need to set ptr to 0?
> Understood, if idleslope and sendslope are 0, then ptr could be set to any
> value and the result would be the same.  And, based on my limited
> understanding, 0 does not seem to be a bad choice.
>
> My point is that ptr shouldn't be uninitialised at this point.

OK, I have sent the v2 version

https://patchwork.kernel.org/project/netdevbpf/patch/20240617013922.1035854-1-xiaolei.wang@windriver.com/

thanks

xiaolei

>
> ...
diff mbox series

Patch

diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_tc.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_tc.c
index 1562fbdd0a04..b0fd2d6e525e 100644
--- a/drivers/net/ethernet/stmicro/stmmac/stmmac_tc.c
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_tc.c
@@ -358,24 +358,26 @@  static int tc_setup_cbs(struct stmmac_priv *priv,
 
 	port_transmit_rate_kbps = qopt->idleslope - qopt->sendslope;
 
-	/* Port Transmit Rate and Speed Divider */
-	switch (div_s64(port_transmit_rate_kbps, 1000)) {
-	case SPEED_10000:
-	case SPEED_5000:
-		ptr = 32;
-		break;
-	case SPEED_2500:
-	case SPEED_1000:
-		ptr = 8;
-		break;
-	case SPEED_100:
-		ptr = 4;
-		break;
-	default:
-		netdev_err(priv->dev,
-			   "Invalid portTransmitRate %lld (idleSlope - sendSlope)\n",
-			   port_transmit_rate_kbps);
-		return -EINVAL;
+	if (qopt->enable) {
+		/* Port Transmit Rate and Speed Divider */
+		switch (div_s64(port_transmit_rate_kbps, 1000)) {
+		case SPEED_10000:
+		case SPEED_5000:
+			ptr = 32;
+			break;
+		case SPEED_2500:
+		case SPEED_1000:
+			ptr = 8;
+			break;
+		case SPEED_100:
+			ptr = 4;
+			break;
+		default:
+			netdev_err(priv->dev,
+				   "Invalid portTransmitRate %lld (idleSlope - sendSlope)\n",
+				   port_transmit_rate_kbps);
+			return -EINVAL;
+		}
 	}
 
 	mode_to_use = priv->plat->tx_queues_cfg[queue].mode_to_use;